Big problem with Naive Bayes assignment

Course Q&A Natural Language Processing NLP with Classification and Vector Spaces
1

Hello everyone.
I’ve a big problem with Naive Bayes assignment because in this part

image
image558×213 4.64 KB

Although the procedure is correct, it tells me that it is wrong. I don’t know if there is a possibility to reset and delete everything I did to check if, due to human error, I changed something in the base code.
I appreciate your prompt response.

2

To get a new copy of the assignment and start over:

  • Use the File → Rename menu to rename your notebook.
  • Then use Lab Help and “Get latest version”.
  • Then you can use File → Open to start using the new notebook.

“Lab Help” is the question-mark inside a circle.

1 Like
3

Note that this is NLP Course 1, not NLP Course 2, so I changed the title category for you.

If you still have the same problem after starting again with a fresh notebook using Tom’s instructions, then please let us know.

There are a number of common mistakes here, but I have not previously seen the logprior value being wrong. So it may be worth looking at your code. We can’t do that on a public thread, but there are private ways to do that. Please check your DMs (personal messages) for a mesasge from me about how to proceed.

1 Like
4

Thanks a lot.

5

Thank you very much.

6

Note that the most probable cause here is that your code is actually wrong, so it would be a good first step to read through the instructions carefully. Pay particular attention to the purpose and meaning of the logprior value. The result should be 0.0 here, because the number of positive and negative input tweets is the same. Meaning that there should be no natural positive or negative bias in the training data. You need to figure out why that value is negative in your code.

1 Like
7

To close the loop on the public thread, there was a bug in how logprior was being calculated.

8

Hint: log_e(x) < 0 \implies x \in (0, 1)

2 Likes